Scheduled maintenance is currently in progress. We will provide updates as necessary.
Posted Feb 02, 2023 - 07:00 UTC
Scheduled
Starting on February 1st, the compression codec for Archive Parquet will be changed from SNAPPY to ZSTD.
Due to this change, archive data will be compressed more efficiently, thereby reducing S3 storage usage and S3 costs for customers as well as improving the overall performance of S3 queries.
When querying parquet files directly, customers should be aware that their framework of choice must support ZSTD. It should be enabled by default in most cases and anyway shouldn't affect regular queries, but some additional setup may be required if customers are using external tools to query the S3 archive (for example, Spark users must ensure ZSTD bindings are included in their classpath).
Our team can maintain the old compression format if customers are not ready for the change and are running jobs on that data. In this case please reach out to our support or your CSM and we will gladly help.